Tutorial Exercise 3

Estimating False Discovery Rates in Strategy Development

# Load required libraries
library(tidyverse)
library(ggplot2)
library(knitr)
library(kableExtra)
library(data.table)
library(gridExtra) # For arranging multiple plots

Introduction

In this tutorial, we will explore the problem of false discoveries in quantitative finance, specifically focusing on how multiple testing leads to selection bias in strategy development. We will:

1. Simulate the process of strategy development with random strategies
2. Calculate the impact of selection bias on Sharpe ratios
3. Implement the Deflated Sharpe Ratio (DSR) to correct for selection bias
4. Analyze how false discovery rates change with different parameters

This exercise will provide hands-on experience with the concepts discussed in the lecture on backtest overfitting.

Why False Discovery Rates Matter in Finance

Financial strategy development differs fundamentally from many scientific disciplines:

1. Limited data: Unlike fields with abundant data (e.g., genomics), finance has relatively few independent observations
2. High noise-to-signal ratio: Market returns contain substantial random variation, making pattern detection challenging
3. Non-stationarity: Financial markets evolve over time, invalidating previously discovered relationships
4. Practical consequences: Implementing false strategies leads to real financial losses

When a researcher tests multiple strategies and selects only the best performer, they risk identifying patterns that exist purely by chance. Understanding false discovery rates is essential for distinguishing genuine market inefficiencies from statistical artifacts.

Part 1: Generating Random Strategies

Let’s first create a function to generate random strategy returns. These will simulate strategies with no genuine edge (i.e., the true Sharpe ratio is zero), as well as some strategies with a small genuine edge.

# Function to generate random strategy returns
# This simulates the returns of investment strategies that have no real edge
# Modified function to generate strategies, some with genuine edge
generate_random_strategies <- function(n_strategies = 1000, 
                                       n_returns = 252, 
                                       mean_return = 0, 
                                       sd_return = 0.01,
                                       edge_pct = 0.05,    # Percentage of strategies with edge (typical in finance)
                                       edge_size = 0.0005) # Size of daily edge (~12% annual)
{
  # Input validation
  if(!is.numeric(n_strategies) || n_strategies <= 0) 
    stop("n_strategies must be a positive number")
  if(!is.numeric(n_returns) || n_returns <= 0) 
    stop("n_returns must be a positive number")
  if(!is.numeric(edge_pct) || edge_pct < 0 || edge_pct > 1) 
    stop("edge_pct must be between 0 and 1")
  
  # Create a matrix of random returns (normally distributed)
  returns_matrix <- matrix(
    rnorm(n_strategies * n_returns, mean = mean_return, sd = sd_return),
    nrow = n_returns,
    ncol = n_strategies
  )
  
  # Add edge to a subset of strategies
  n_edge_strategies <- round(n_strategies * edge_pct)
  if (n_edge_strategies > 0) {
    # Add a small positive drift to create genuine edge
    for (i in 1:n_edge_strategies) {
      returns_matrix[, i] <- returns_matrix[, i] + edge_size
    }
  }
  
  # Name each strategy for easier reference
  colnames(returns_matrix) <- paste0("Strategy_", 1:n_strategies)
  
  # Create a vector indicating which strategies have true edge
  has_edge <- logical(n_strategies)
  has_edge[1:n_edge_strategies] <- TRUE
  
  return(list(
    returns = returns_matrix,
    has_edge = has_edge
  ))
}

# Generate strategies with some having genuine edge
set.seed(42)
strategy_data <- generate_random_strategies(
  n_strategies = 1000,
  n_returns = 252,
  edge_pct = 0.05,    # 5% of strategies have genuine edge
  edge_size = 0.0005  # Small daily drift (about 12% annual return)
)

# Extract the returns matrix and edge information
random_strategies <- strategy_data$returns
true_edge <- strategy_data$has_edge

# Check the dimensions to confirm we have 252 days × 1000 strategies
dim(random_strategies)

[1]  252 1000

Important

Instructor Note: We use simplified assumptions (normal, IID returns) for clarity and teaching effectiveness. Real financial markets exhibit additional complexity (e.g., skewness, kurtosis, volatility clustering), which you might explore in more advanced exercises or projects.

library(sn) # Skew-normal package

generate_random_strategies <- function(n_strategies = 1000, 
                                       n_returns = 252, 
                                       mean_return = 0, 
                                       sd_return = 0.01,
                                       skewness = 0,      # set to non-zero for skewness
                                       edge_pct = 0.05, 
                                       edge_size = 0.0005) {
  returns_matrix <- matrix(
    rsn(n_strategies * n_returns, xi = mean_return, omega = sd_return, alpha = skewness),
    nrow = n_returns,
    ncol = n_strategies
  )
  
  # Add edge to a subset of strategies
  n_edge_strategies <- round(n_strategies * edge_pct)
  if (n_edge_strategies > 0) {
    returns_matrix[, 1:n_edge_strategies] <- returns_matrix[, 1:n_edge_strategies] + edge_size
  }
  
  colnames(returns_matrix) <- paste0("Strategy_", 1:n_strategies)
  has_edge <- logical(n_strategies)
  has_edge[1:n_edge_strategies] <- TRUE
  
  return(list(returns = returns_matrix, has_edge = has_edge))
}

library(rugarch)

generate_garch_strategies <- function(n_strategies = 1000, n_returns = 252, edge_pct = 0.05, edge_size = 0.0005) {
  spec <- ugarchspec(mean.model = list(armaOrder = c(0,0)),
                     variance.model = list(model = "sGARCH", garchOrder = c(1,1)),
                     distribution.model = "norm")
  
  returns_matrix <- sapply(1:n_strategies, function(x) {
    sim <- ugarchpath(spec, n.sim = n_returns)
    fitted_returns <- as.numeric(fitted(sim))
    return(fitted_returns / sd(fitted_returns) * 0.01) # standardized volatility
  })
  
  # Add edge to a subset of strategies
  n_edge_strategies <- round(n_strategies * edge_pct)
  if (n_edge_strategies > 0) {
    returns_matrix[, 1:n_edge_strategies] <- returns_matrix[, 1:n_edge_strategies] + edge_size
  }

  colnames(returns_matrix) <- paste0("Strategy_", 1:n_strategies)
  has_edge <- logical(n_strategies)
  has_edge[1:n_edge_strategies] <- TRUE
  
  return(list(returns = returns_matrix, has_edge = has_edge))
}

Now, let’s calculate the Sharpe ratio for each of these strategies.

# Function to calculate NON-annualized Sharpe ratio
# Useful for statistical tests (e.g., Deflated Sharpe Ratio)
calculate_sharpe <- function(returns) {
  sd_return <- sd(returns)
  if (sd_return == 0) return(0)
  mean(returns) / sd_return
}

# Function to calculate annualized Sharpe ratio
# Typically used in industry to report performance
calculate_sharpe_annualized <- function(returns, annualization_factor = 252) {
  calculate_sharpe(returns) * sqrt(annualization_factor)
}

# Compute non-annualized Sharpe ratios for DSR and statistical testing
sharpe_ratios <- apply(random_strategies, 2, calculate_sharpe)

# Compute annualized Sharpe ratios for reporting and visualization
sharpe_ratios_annualized <- sharpe_ratios * sqrt(252)

# Examine distribution statistics of annualized Sharpe ratios
summary(sharpe_ratios_annualized)

    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
-3.12341 -0.63281 -0.00652  0.02047  0.69419  3.22375 

# Separate Sharpe ratios (annualized) for strategies with and without edge
edge_sharpes <- sharpe_ratios_annualized[true_edge]
no_edge_sharpes <- sharpe_ratios_annualized[!true_edge]

# Create data frame clearly labeling annualized Sharpe ratios for plotting
sharpe_df <- data.frame(
  Annualized_Sharpe = sharpe_ratios_annualized,
  Has_Edge = factor(true_edge, levels = c(FALSE, TRUE),
                    labels = c("No Edge", "True Edge"))
)

# Visualize clearly labeled Annualized Sharpe ratios
ggplot(sharpe_df, aes(x = Annualized_Sharpe, fill = Has_Edge)) +
  geom_histogram(bins = 30, alpha = 0.7, position = "identity") +
  labs(
    title = "Distribution of Annualized Sharpe Ratios for Random Strategies",
    subtitle = paste("Mean =", round(mean(sharpe_ratios_annualized), 2), 
                     "SD =", round(sd(sharpe_ratios_annualized), 2),
                     "| 5% of strategies have true edge"),
    x = "Annualized Sharpe Ratio",
    y = "Count",
    fill = "Strategy Type"
  ) +
  theme_minimal() +
  geom_vline(xintercept = 0, linetype = "dashed", color = "black") +
  scale_fill_manual(values = c("No Edge" = "steelblue", "True Edge" = "darkred"))

The histogram above shows the distribution of Sharpe ratios for our 1,000 randomly generated strategies. This is a critical visualization that illustrates several important concepts in quantitative finance:

Important Clarification on Sharpe Ratio:

The Sharpe ratio can be expressed either in non-annualized or annualized form.

• Non-annualized Sharpe ratio is used in statistical computations such as the Deflated Sharpe Ratio, as it directly reflects the actual daily returns and volatility without scaling.
• Annualized Sharpe ratio is typically reported in practice to standardise performance to a common time frame (usually one year).

In this tutorial, we explicitly distinguish between the two to ensure accurate statistical analysis (DSR) while maintaining practical relevance for visualizations and reporting.

Key Observations:

1. Normal Distribution: Notice how the Sharpe ratios for strategies without edge follow an approximately normal distribution, centered near zero. This is exactly what we would expect when strategies have no genuine edge – their performance is purely random.

2. Standard Deviation: The standard deviation of Sharpe ratios tells us about the spread of performance metrics. This is crucial because it determines how impressive a “good” Sharpe ratio needs to be to stand out from random noise.

3. Range of Values: Even though most strategies are completely random (with no edge), we see Sharpe ratios ranging from approximately -3 to +3. In practical terms, a Sharpe ratio of +2 is typically considered excellent in the investment industry! Yet here we see several random strategies achieving this level purely by chance.

4. Strategies with Edge: The strategies with genuine edge (in dark red) tend to have higher Sharpe ratios on average, but there’s substantial overlap with the no-edge strategies. This illustrates why it’s so difficult to identify genuine strategies from random ones based on Sharpe ratio alone.

Implications for Strategy Development:

This distribution demonstrates why statistical significance is so important in strategy evaluation. If you tested just one strategy and it showed a Sharpe ratio of 1.5, you might be excited about its performance. However, this chart shows that among 1,000 random strategies, we’d expect several to show Sharpe ratios of 1.5 or higher simply due to chance.

This is precisely why multiple testing is so dangerous in quantitative finance. When researchers or traders test many strategy variations and report only the best performer, they are essentially “selecting” from the right tail of this distribution – capturing lucky outcomes rather than genuine edge.

In the next sections, we’ll explore exactly how the maximum Sharpe ratio increases with the number of trials, and how we can use the Deflated Sharpe Ratio to correct for this selection bias.

Checkpoint Questions: Understanding Random Strategy Performance

Take a moment to answer these questions:

1. What key observation can we make about the distribution of Sharpe ratios from our random strategies?
2. Why did some random strategies show Sharpe ratios above 2.0 despite having no true edge?
3. If you developed a strategy with a Sharpe ratio of 1.8, how would you assess whether it was likely to be genuine?

Click for answers

1. The Sharpe ratios follow an approximately normal distribution centered near zero (mean = -0.02), with a standard deviation of approximately 0.98.
2. With 1,000 random strategies, we expect to see some extremely high Sharpe ratios purely by chance - this is a statistical property of multiple testing.
3. You should consider how many alternative strategies were tested, estimate the expected maximum Sharpe ratio under the null hypothesis of no edge, and calculate the Deflated Sharpe Ratio to assess the probability that your 1.8 Sharpe is not simply the result of selection bias.

Part 2: Exploring Selection Bias

Now, let’s simulate the process of strategy selection, where a researcher tests multiple strategies and selects the one with the highest Sharpe ratio.

# Find the maximum Sharpe ratio from our 1000 strategies
# This simulates what happens when a researcher selects only the best-performing strategy
max_sharpe <- max(sharpe_ratios_annualized)
max_sharpe_index <- which.max(sharpe_ratios_annualized)

# Check if the best-performing strategy has true edge
has_true_edge <- true_edge[max_sharpe_index]

# Print the maximum Sharpe ratio and which strategy achieved it
# This is often what would be presented in a backtest report or research paper
cat("Maximum Sharpe Ratio:", round(max_sharpe, 2), 
    "\nStrategy Index:", max_sharpe_index, 
    "\nHas True Edge:", has_true_edge, "\n")

Maximum Sharpe Ratio: 3.22 
Strategy Index: 851 
Has True Edge: FALSE 

# Define a function to calculate the expected maximum Sharpe ratio
# This implements the False Strategy Theorem from the lecture
# It predicts how high the max Sharpe should be just from random chance
expected_max_sharpe <- function(n_trials, mean_sr = 0, std_sr = 1) {
  # Euler-Mascheroni constant (mathematical constant appearing in the theorem)
  emc <- 0.57721566490153286060651209008240243104215933593992
  
  # Expected maximum Sharpe ratio formula:
  # The more trials we run, the higher this value becomes, even with random data
  sr0 <- (1 - emc) * qnorm(p = 1 - 1/n_trials) + 
        emc * qnorm(1 - (n_trials * exp(1))^(-1))
  
  # Adjust by the mean and standard deviation of the Sharpe ratio distribution
  sr0 <- mean_sr + std_sr * sr0
  
  return(sr0)
}

# Calculate the expected maximum Sharpe ratio for 1000 trials
# We use the actual standard deviation of our Sharpe ratios for accuracy
exp_max_sharpe <- expected_max_sharpe(1000, mean_sr = 0, std_sr = sd(sharpe_ratios_annualized))

# Compare the theoretical expectation with our actual observed maximum
# They should be relatively close if our strategies are truly random
cat("Expected Maximum Sharpe Ratio:", round(exp_max_sharpe, 2), "\n")

Expected Maximum Sharpe Ratio: 3.23 

cat("Actual Maximum Sharpe Ratio:", round(max_sharpe, 2), "\n")

Actual Maximum Sharpe Ratio: 3.22 

Let’s visualise how the maximum Sharpe ratio increases with the number of trials.

# Calculate expected maximum Sharpe ratio for different numbers of trials
# This helps us visualize how the "best strategy" improves just by testing more variations
trial_counts <- c(1, 5, 10, 50, 100, 500, 1000)
exp_max_sharpes <- sapply(trial_counts, expected_max_sharpe, 
                          mean_sr = 0, std_sr = sd(sharpe_ratios_annualized))

# Function to simulate actual maximum Sharpe ratios for different trial counts
# This gives us an empirical check against the theoretical prediction
simulate_max_sharpe <- function(n_trials, n_simulations = 100, 
                               mean_sr = 0, std_sr = 1) {
  # We'll run multiple simulations to get a distribution of maximum Sharpe ratios
  max_sharpes <- numeric(n_simulations)
  
  # For each simulation, generate n_trials random Sharpe ratios and find the maximum
  for (i in 1:n_simulations) {
    # Generate random Sharpe ratios with specified mean and std
    sharpes <- rnorm(n_trials, mean = mean_sr, sd = std_sr)
    # Record the maximum value - this simulates selecting the best strategy
    max_sharpes[i] <- max(sharpes)
  }
  
  # Return statistics about the distribution of maximum Sharpe ratios
  return(list(
    mean = mean(max_sharpes),  # Average maximum across simulations
    sd = sd(max_sharpes),      # Standard deviation of the maximums
    values = max_sharpes       # All the individual maximum values
  ))
}

# Run the simulation for each trial count
# For each number of trials, we simulate 100 sets to get a good estimate
set.seed(123)  # For reproducibility
simulated_results <- lapply(trial_counts, simulate_max_sharpe, 
                           n_simulations = 100, 
                           mean_sr = 0, 
                           std_sr = sd(sharpe_ratios_annualized))

# Extract the mean maximum Sharpe ratio from each simulation set
sim_max_sharpes <- sapply(simulated_results, function(x) x$mean)

# Create a data frame for plotting both theoretical and simulated results
plot_data <- tibble(
  Trials = rep(trial_counts, 2),  # Each trial count appears twice (theoretical and simulated)
  `Sharpe Ratio` = c(exp_max_sharpes, sim_max_sharpes),  # Values from both methods
  Type = rep(c("Theoretical", "Simulated"), each = length(trial_counts))  # Identify the source
)

# Create the plot comparing theoretical vs. simulated maximum Sharpe ratios
# This is a key visualization demonstrating how selection bias increases with trials
ggplot(plot_data, aes(x = Trials, y = `Sharpe Ratio`, color = Type)) +
  geom_line(size = 1) +
  geom_point(size = 3) +
  scale_x_log10() +  # Log scale makes the pattern clearer across different trial counts
  labs(
    title = "Maximum Sharpe Ratio vs. Number of Trials",
    subtitle = "Theoretical (False Strategy Theorem) vs. Simulated",
    x = "Number of Trials (log scale)",
    y = "Maximum Sharpe Ratio"
  ) +
  theme_minimal() +
  scale_color_manual(values = c("Theoretical" = "darkblue", "Simulated" = "darkred"))

The results above demonstrate one of the most important concepts in quantitative finance: the expected maximum Sharpe ratio increases systematically with the number of trials, even when all strategies have zero true edge.

Looking at our data:

1. With just a single random strategy (trials = 1), the expected maximum Sharpe ratio is close to zero.
2. With 10 trials, we expect a maximum Sharpe around 1.5.
3. With 100 trials, the expected maximum rises to about 2.3.
4. With 1000 trials, we expect a maximum Sharpe ratio above 3.0!

This relationship is both theoretical (as predicted by the False Strategy Theorem) and empirical (as shown by our simulations). The close match between our theoretical and simulated lines confirms the validity of the theorem.

The implications for research are profound:

• A strategy with a Sharpe ratio of 2.0 might seem impressive in isolation
• But if it was selected as the best out of 100+ configurations tested, it’s actually exactly what we’d expect from random chance
• This explains why so many strategies that look excellent in backtests fail in live trading

This is why proper statistical adjustments like the Deflated Sharpe Ratio (which we’ll explore next) are essential when evaluating investment strategies that resulted from backtesting multiple configurations.

The Selection Bias Visual

Let’s visualise the selection bias problem more directly:

# Generate a set of Sharpe ratios for 100 random strategies
set.seed(456)
random_sharpes <- rnorm(100, mean = 0, sd = 1)

# Create a data frame for visualization
selection_bias_df <- data.frame(
  Strategy = 1:100,
  Sharpe = random_sharpes
)

# Plot the Sharpe ratios with the maximum highlighted
ggplot(selection_bias_df, aes(x = Strategy, y = Sharpe)) +
  geom_hline(yintercept = 0, linetype = "dashed", color = "gray") +
  geom_segment(aes(xend = Strategy, yend = 0), color = "steelblue", alpha = 0.5) +
  geom_point(color = "steelblue", size = 2) +
  geom_point(data = selection_bias_df[which.max(random_sharpes),], 
             color = "red", size = 4) +
  annotate("text", x = which.max(random_sharpes), 
           y = max(random_sharpes) + 0.3, 
           label = paste("Max SR =", round(max(random_sharpes), 2)), 
           color = "red") +
  labs(
    title = "Selection Bias in Strategy Development",
    subtitle = "100 random strategies with no true edge",
    x = "Strategy Number",
    y = "Sharpe Ratio"
  ) +
  theme_minimal()

This visualisation shows exactly what happens in strategy selection: we run many trials and pick the best one. The strategy with the highest Sharpe ratio (highlighted in red) looks impressive, but it’s purely a result of random chance. If we were to implement this strategy, we would likely be disappointed by its future performance.

Part 3: Calculating the Deflated Sharpe Ratio

Now, let’s implement the Deflated Sharpe Ratio to correct for selection bias under multiple testing.

# Function to calculate skewness and excess kurtosis
calculate_moments <- function(returns) {
  z <- (returns - mean(returns)) / sd(returns)
  skew <- mean(z^3)
  kurt <- mean(z^4) - 3
  return(list(skewness = skew, kurtosis = kurt))
}

# Function to calculate expected maximum Sharpe ratio (False Strategy Theorem)
expected_max_sharpe <- function(n_trials, mean_sr = 0, std_sr = 1) {
  emc <- 0.577215664901532860606512090082402431042159336  # Euler-Mascheroni constant
  sr0 <- (1 - emc) * qnorm(1 - 1 / n_trials) + emc * qnorm(1 - (n_trials * exp(1))^(-1))
  return(mean_sr + std_sr * sr0)
}

calculate_dsr <- function(returns, n_trials, sr_mean = 0, sr_std = NULL, annual_factor = 252) {
  
  n <- length(returns)
  # Compute daily Sharpe ratio explicitly
  sr_daily <- mean(returns) / sd(returns)
  
  # Explicit annualisation
  sr_annualised <- sr_daily * sqrt(annual_factor)
  
  # If sr_std not provided, issue warning and default sensibly
  if (is.null(sr_std) || sr_std <= 0) {
    sr_std <- 1
    warning("SR standard deviation not provided or invalid; using default value of 1 (annualised)")
  }
  
  moments <- calculate_moments(returns)
  skew <- moments$skewness
  kurt <- moments$kurtosis
  
  # Use annualised expected maximum Sharpe ratio consistently
  exp_max_sr_annualised <- expected_max_sharpe(n_trials, mean_sr = sr_mean, std_sr = sr_std)
  
  numerator <- (sr_annualised - exp_max_sr_annualised) * sqrt(n - 1)
  denominator_term <- 1 - skew * sr_daily + ((kurt + 2) / 4) * sr_daily^2
  
  if (denominator_term <= 0 || is.na(denominator_term)) {
    warning("Instability in denominator calculation; returning NA for DSR")
    return(list(
      sharpe_ratio = sr_annualised,
      expected_max_sr = exp_max_sr_annualised,
      dsr = NA,
      skewness = skew,
      kurtosis = kurt,
      n_observations = n
    ))
  } else {
    denominator <- sqrt(denominator_term)
  }
  
  dsr <- pnorm(numerator / denominator)
  
  return(list(
    sharpe_ratio = sr_annualised,
    expected_max_sr = exp_max_sr_annualised,
    dsr = dsr,
    skewness = skew,
    kurtosis = kurt,
    n_observations = n
  ))
}

# Visualize the DSR calculation process
library(DiagrammeR)

# Create flow diagram of DSR calculation process
grViz("
digraph DSR_calculation {
  # Node definitions
  node [shape = rectangle, style = filled, fillcolor = lightblue, fontname = Helvetica]
  
  A [label = 'Input: Strategy returns\\nand number of trials']
  B [label = 'Calculate Sharpe ratio\\n(SR = mean/std)']
  C [label = 'Calculate higher moments\\n(skewness, kurtosis)']
  D [label = 'Estimate expected maximum SR\\n(False Strategy Theorem)']
  E [label = 'Calculate DSR\\n(probability measure)']
  F [label = 'Interpret DSR result\\n(probability of true discovery)']
  
  # Edge definitions
  A -> B
  B -> C
  {B; C} -> D
  {B; C; D} -> E
  E -> F
  
  # Graph attributes
  graph [rankdir = TB, splines = true, nodesep = 0.8]
}
")

# Calculate DSR for our maximum Sharpe ratio strategy
# This applies our DSR function to the strategy that looked best in the backtest
best_strategy_returns <- random_strategies[, max_sharpe_index]
dsr_result <- calculate_dsr(
  best_strategy_returns, 
  n_trials = 1000,  # We tested 1000 strategies
  sr_std = sd(sharpe_ratios_annualized)  # Use the actual variability in our Sharpe ratios
)

# Print the results
# These metrics tell us whether our "best strategy" is likely genuine or just lucky
cat("Sharpe Ratio:", round(dsr_result$sharpe_ratio, 2), "(annualized)\n")

Sharpe Ratio: 3.22 (annualized)

cat("Expected Max Sharpe Ratio:", round(dsr_result$expected_max_sr, 2), "\n")

Expected Max Sharpe Ratio: 3.23 

cat("Deflated Sharpe Ratio:", round(dsr_result$dsr, 4), "\n")

Deflated Sharpe Ratio: 0.4428 

cat("Skewness:", round(dsr_result$skewness, 2), "\n")

Skewness: -0.16 

cat("Excess Kurtosis:", round(dsr_result$kurtosis, 2), "\n")

Excess Kurtosis: -0.27 

Understanding the Deflated Sharpe Ratio Result

The Deflated Sharpe Ratio (DSR) we’ve calculated represents the probability that our strategy’s performance is not merely due to selection bias. Let’s interpret our results:

1. Sharpe Ratio: Our best strategy has an annualized Sharpe ratio of 3.22. In traditional finance, this would be considered excellent performance, well above the typical threshold of 1.0 for investment consideration.

2. Expected Maximum Sharpe Ratio: However, given that we tested 1000 strategies, we would expect to find a maximum Sharpe ratio of approximately 3.20 purely by chance. This is a critical benchmark that any truly successful strategy must exceed.

3. Deflated Sharpe Ratio: The DSR value of 0.44 means there’s approximately a 44% probability that this high Sharpe ratio is genuinely indicative of a strategy’s true predictive power, rather than being merely due to luck from multiple tests. In practice, this means the strategy’s validity is quite uncertain—neither clearly false nor clearly genuine.

4. Higher Moments (Skewness & Kurtosis): With skewness of -0.16 and excess kurtosis of -0.27, our returns distribution closely approximates normality, indicating relatively stable strategy returns without extreme outliers. This reduces, though does not eliminate, uncertainty from extreme events influencing our Sharpe ratio.

Industry Decision Framework

In professional quantitative investment firms, DSR thresholds often guide strategy implementation decisions:

DSR Range	Interpretation	Typical Action
0.00-0.20	Strong evidence of false discovery	Reject strategy
0.20-0.50	Likely false discovery or uncertain	Consider additional testing or validation on new data before implementation
0.50-0.80	Uncertain (moderate evidence)	Additional testing with independent data required
0.80-0.95	Potentially valid (good evidence)	Implement with caution, possibly at reduced scale
0.95-1.00	Strong evidence of valid strategy	Full implementation consideration

With a DSR of 0.44, your current strategy would sit firmly in the “Likely false or uncertain” category, demanding careful further validation.

In the next section, we’ll explore how different parameters affect the false discovery rate, which will further illustrate why traditional backtesting approaches are so problematic.

Part 4: False Discovery Rate Analysis

Let’s analyze how the false discovery rate changes with different parameters.

# Function to calculate precision, recall, and False Discovery Rate (FDR)
# This implements the mathematical framework from Lopez de Prado discussed in the lecture
calculate_fdr <- function(ground_truth, alpha = 0.05, beta = 0.2) {
  # Input validation
  if(ground_truth < 0 || ground_truth > 1) 
    stop("ground_truth must be between 0 and 1")
  if(alpha < 0 || alpha > 1) 
    stop("alpha must be between 0 and 1")
  if(beta < 0 || beta > 1) 
    stop("beta must be between 0 and 1")
  
  # Convert ground truth probability to odds ratio (theta)
  # ground_truth = proportion of strategies that are truly profitable
  # theta = ratio of true strategies to false strategies
  theta <- ground_truth / (1 - ground_truth)
  
  # Calculate recall (true positive rate)
  # Recall = 1 - beta, where beta is the Type II error rate (false negative rate)
  # This represents the probability of detecting a true strategy
  recall <- 1 - beta          
  
  # Calculate numerator for precision calculation
  # b1 = recall * theta = number of true positives / number of false strategies
  b1 <- recall * theta
  
  # Calculate precision
  # Precision = true positives / all positives
  # This represents the probability that a positive test indicates a true strategy
  precision <- b1 / (b1 + alpha)
  
  # Calculate False Discovery Rate (FDR)
  # FDR = 1 - precision = false positives / all positives
  # This is the probability that a strategy that tests positive is actually false
  fdr <- 1 - precision
  
  # Return all relevant metrics in a tidy format
  return(tibble(
    ground_truth = ground_truth,  # Original input - prior probability of true strategy
    theta = theta,                # Odds ratio of true vs. false strategies
    alpha = alpha,                # Type I error rate (significance level)
    beta = beta,                  # Type II error rate (1 - power)
    recall = recall,              # True positive rate (power)
    precision = precision,        # Proportion of positives that are true
    fdr = fdr                     # Proportion of positives that are false
  ))
}

# Calculate FDR for different ground truth probabilities
# This shows how the FDR changes based on the prior probability of true strategies
ground_truths <- seq(0.01, 0.5, by = 0.01)  # Try values from 1% to 50%
fdr_results <- map_df(ground_truths, calculate_fdr)  # Apply function to each value

# Plot the results to visualize the relationship
# This is a key insight: even with standard statistical testing, FDR remains high
# when true strategies are rare (which is the case in finance)
ggplot(fdr_results, aes(x = ground_truth)) +
  geom_line(aes(y = precision, color = "Precision"), size = 1) +
  geom_line(aes(y = fdr, color = "FDR"), size = 1) +
  scale_color_manual(values = c("Precision" = "blue", "FDR" = "red")) +
  labs(
    title = "Precision and False Discovery Rate vs. Ground Truth Probability",
    subtitle = "Alpha = 0.05, Beta = 0.2",
    x = "Ground Truth Probability (Proportion of True Strategies)",
    y = "Rate",
    color = "Metric"
  ) +
  theme_minimal() +
  scale_x_continuous(labels = scales::percent) +  # Format x-axis as percentages
  scale_y_continuous(labels = scales::percent) +   # Format y-axis as percentages
  # Add a vertical line at the point where precision = 0.5 (FDR = 0.5)
  geom_vline(xintercept = 0.0625, linetype = "dashed", color = "black") +
  annotate("text", x = 0.09, y = 0.5, 
           label = "Precision = 50%\nwhen ground truth ≈ 6.25%", 
           hjust = 0)

Now, let’s see how the FDR changes with different significance levels (alpha).

# Calculate FDR for different alpha values (significance levels)
alphas <- seq(0.01, 0.1, by = 0.01)  # Test alpha values from 1% to 10%

# Create combinations of ground truth probabilities and alphas
fdr_by_alpha <- expand.grid(
  initial_ground_truth = c(0.01, 0.05, 0.1, 0.2),  # Different prior probabilities
  initial_alpha = alphas                           # Different significance levels
) %>%
  as_tibble() %>%
  rowwise() %>%
  # Calculate FDR for each combination
  mutate(
    result = list(calculate_fdr(initial_ground_truth, initial_alpha, beta = 0.2))
  ) %>%
  unnest(result) %>%
  # Use the values from the calculate_fdr result, dropping the duplicates
  select(-initial_ground_truth, -initial_alpha)

# Plot FDR vs alpha for different ground truth probabilities
ggplot(fdr_by_alpha, aes(x = alpha, y = fdr, color = factor(ground_truth))) +
  geom_line(size = 1) +
  labs(
    title = "False Discovery Rate vs. Significance Level",
    subtitle = "For Different Ground Truth Probabilities",
    x = "Significance Level (Alpha)",
    y = "False Discovery Rate",
    color = "Ground Truth"
  ) +
  theme_minimal() +
  scale_y_continuous(labels = scales::percent) +
  scale_color_discrete(name = "Ground Truth", 
                      labels = scales::percent(unique(fdr_by_alpha$ground_truth)))

Understanding False Discovery Rates in Finance

The plots above reveal a profound challenge in quantitative finance - the false discovery rate is alarmingly high even with traditional statistical safeguards.

Key Insights from the First Plot:

1. High False Discovery Rate with Low Prior Probability: When only 1% of tested strategies are genuinely profitable (a realistic scenario in finance), the false discovery rate is approximately 94% even when using the standard significance level of 5%. This means that 94% of “discoveries” are actually false!

2. The Prior Probability Problem: Notice how dramatically the FDR drops as the ground truth probability increases. In fields like medicine or physics, where the prior probability of a true effect might be 20-50%, traditional statistical methods work well. But in finance, where market efficiency makes true strategies rare, these methods break down.

3. The Reliability Threshold: The ground truth probability needs to exceed 6.25% before the false discovery rate drops below 50% (with α=0.05 and β=0.2). This means that without a strong prior belief in a strategy’s effectiveness, most statistically significant backtests are likely to be false.

Key Insights from the Second Plot:

1. Significance Level Impact: Making our statistical tests more stringent (lower α) does reduce the false discovery rate, but the effect is modest compared to the impact of the ground truth probability.

2. Diminishing Returns: Even with a very strict significance level of α=0.01, the false discovery rate remains above 80% when the ground truth probability is just 1%.

3. The Multiple Testing Connection: This analysis reveals why multiple testing is so problematic in finance - each test increases the opportunity for false positives in an environment where true positives are rare.

These findings help explain why so many published financial strategies fail to replicate and why institutional investors are rightfully skeptical of backtested performance. They also underscore the importance of methods like the Deflated Sharpe Ratio, which explicitly account for multiple testing and the low prior probability of true strategies.

Checkpoint Questions: False Discovery Rates

Take a moment to reflect on these questions:

1. Why is the false discovery rate so high in finance compared to other fields?
2. If you were examining a strategy with a statistically significant p-value of 0.01, what other information would you need to assess its true validity?
3. How might you increase the prior probability (ground truth) of your strategies before backtesting?

Click for answers

1. The false discovery rate is high in finance because the prior probability of true strategies (ground truth) is typically very low due to market efficiency. Even with strict statistical tests, if the base rate of true strategies is low, most “discoveries” will be false positives.

2. You would need to know: (a) how many strategies were tested before finding this one (to assess selection bias), (b) the prior probability of true strategies in your domain, and (c) the power of your test (1-β). With this information, you could calculate the true probability that your strategy is a false discovery.

3. To increase prior probability, you could: (a) develop strategies based on strong economic rationales rather than data mining, (b) focus on areas of the market with known inefficiencies, (c) incorporate non-public information (where legally permitted), and (d) apply domain expertise to filter strategy ideas before backtesting.

Part 5: Impact of Sample Size on DSR

Let’s investigate how the sample size affects the Deflated Sharpe Ratio.

# Generate random strategies with different sample sizes
# This helps us understand how the amount of data affects our ability to detect true strategies
sample_sizes <- c(63, 126, 252, 504, 1008)  # Approx. 3 months to 4 years of daily data

# Corrected function to calculate DSR for different sample sizes
calculate_dsr_by_sample <- function(sample_size, n_strategies = 100, n_trials = 100,
                                    edge_pct = 0.05, edge_size = 0.0005) {
  # Generate strategies
  strategies_data <- generate_random_strategies(
    n_strategies = n_strategies,
    n_returns = sample_size,
    edge_pct = edge_pct,
    edge_size = edge_size
  )
  
  strategies <- strategies_data$returns
  has_edge <- strategies_data$has_edge
  
  # Calculate Sharpe ratios (annualised)
  sharpes <- apply(strategies, 2, calculate_sharpe_annualized, annualization_factor = 252)
  
  # Find max Sharpe and its index
  max_sharpe <- max(sharpes)
  max_idx <- which.max(sharpes)
  
  # Check if max strategy has true edge
  best_has_edge <- has_edge[max_idx]
  
  # Calculate DSR for best strategy with correct parameters
  dsr_result <- tryCatch({
    calculate_dsr(
      strategies[, max_idx],
      n_trials = n_trials,
      sr_std = sd(sharpes)
    )
  }, error = function(e) {
    warning("Error in DSR calculation: ", e$message)
    return(NULL)
  })
  
  if (is.null(dsr_result)) {
    return(tibble(
      sample_size = sample_size,
      sharpe_ratio = NA_real_,
      expected_max_sr = NA_real_,
      dsr = NA_real_,
      has_edge = best_has_edge
    ))
  } else {
    return(tibble(
      sample_size = sample_size,
      sharpe_ratio = dsr_result$sharpe_ratio,  # Corrected: already annualised
      expected_max_sr = dsr_result$expected_max_sr,
      dsr = dsr_result$dsr,
      has_edge = best_has_edge
    ))
  }
}

# Run multiple simulations consistently (e.g., 100 strategies, 100 trials)
set.seed(456)
n_simulations <- 50
dsr_by_sample <- map_df(sample_sizes, function(sample_size) {
  map_df(1:n_simulations, function(i) {
    calculate_dsr_by_sample(sample_size, n_strategies = 100, n_trials = 100)
  })
})

library(dplyr)

library(ggridges)
ggplot(dsr_by_sample, aes(x = factor(sample_size), y = dsr, fill = has_edge)) +
  geom_violin(trim = FALSE, alpha = 0.6, position = position_dodge(width = 0.8)) +
  geom_jitter(aes(colour = has_edge), width = 0.15, alpha = 0.8, size = 1, show.legend = FALSE) +
  labs(
    title = "Violin Plot of DSR",
    subtitle = "True vs False Strategies by Sample Size",
    x = "Sample Size (Days)",
    y = "Deflated Sharpe Ratio (DSR)",
    fill = "Strategy Has True Edge"
  ) +
  theme_minimal() +
  scale_fill_manual(values = c("FALSE" = "lightblue", "TRUE" = "darkred")) +
  scale_colour_manual(values = c("FALSE" = "steelblue", "TRUE" = "darkred"))

# # Plot results clearly
# ggplot(dsr_by_sample, aes(x = factor(sample_size), y = dsr, fill = has_edge)) +
#   geom_boxplot() +
#   labs(
#     title = "Distribution of Deflated Sharpe Ratio by Sample Size",
#     subtitle = "100 Strategies, 100 Trials, 50 Simulations per Sample Size",
#     x = "Sample Size (Number of Returns)",
#     y = "Deflated Sharpe Ratio",
#     fill = "Strategy Has\nTrue Edge"
#   ) +
#   theme_minimal() +
#   scale_fill_manual(values = c("FALSE" = "lightblue", "TRUE" = "darkred"))

The Critical Impact of Sample Size on Strategy Evaluation

The violin plot visualization above provides critical insights into how sample size significantly influences our ability to accurately distinguish strategies with genuine predictive power (“true edge”) from false discoveries:

1. High Variability of DSR at Small Sample Sizes

At smaller sample sizes (e.g., 63 days, approximately three months), the Deflated Sharpe Ratio (DSR) values exhibit substantial variability, illustrated by the wide distribution in the violin plots. This variability implies that short backtests can easily produce misleading conclusions about strategy quality, leading investors to mistakenly trust strategies that appear statistically sound but are in reality just fortunate anomalies.

2. Improved Stability and Clarity with Increasing Sample Size

As the sample size increases (moving from left to right across the violin plots), the variability of the DSR diminishes noticeably. By the time we reach one or more years of daily data (252 to 1008 days), the distributions become narrower and more clearly differentiated. This improvement reflects greater statistical precision, enabling more reliable assessments of strategy validity.

3. Enhanced Ability to Distinguish True Strategies

Strategies with true edges (highlighted in dark red) tend to maintain consistently higher DSR values as sample size grows. In contrast, false strategies (light blue) become increasingly concentrated around lower DSR values with larger sample sizes. This pattern demonstrates that the reliability of the DSR as a tool for discriminating genuinely effective strategies from those resulting from random chance improves significantly with more data.

4. Practical Implications for Backtesting Practices

This analysis clearly supports industry best practices requiring extensive backtest periods:

• Short periods (e.g., three months) lead to highly unreliable DSR assessments.
• Intermediate periods (e.g., one year) start providing meaningful, albeit imperfect, differentiation.
• Longer periods (e.g., four years) substantially enhance confidence that identified strategies are likely genuine.

5. Balancing Statistical Confidence and Market Relevance

While longer backtests enhance statistical reliability, they inherently incorporate more diverse market conditions, potentially introducing regime shifts and structural changes. Thus, a long sample period can sometimes mask recent market dynamics or structural breaks. Investors must therefore balance rigorous statistical testing against the need to reflect current market conditions realistically.

In summary, the violin plot clearly illustrates that larger sample sizes significantly reduce the risk of false discoveries. Yet, the choice of sample length should thoughtfully incorporate statistical considerations alongside economic and practical market insights.

Statistical Significance vs. Strategy Stability

Let’s examine the relationship between statistical significance (Sharpe ratio), DSR, and sample size:

# Calculate average DSR for strategies with and without edge across sample sizes
avg_dsr_by_sample <- dsr_by_sample %>%
  group_by(sample_size, has_edge) %>%
  summarize(
    avg_dsr = mean(dsr, na.rm = TRUE),
    avg_sharpe = mean(sharpe_ratio, na.rm = TRUE),
    .groups = 'drop'
  )

# Plot the average DSR and Sharpe ratio by sample size
p1 <- ggplot(avg_dsr_by_sample, aes(x = factor(sample_size), y = avg_dsr, 
                               color = has_edge, group = has_edge)) +
  geom_line(size = 1) +
  geom_point(size = 3) +
  labs(
    title = "Average DSR by Sample Size",
    x = "Sample Size (Number of Returns)",
    y = "Average DSR",
    color = "Strategy Has\nTrue Edge"
  ) +
  theme_minimal() +
  scale_color_manual(values = c("FALSE" = "blue", "TRUE" = "red"))

p2 <- ggplot(avg_dsr_by_sample, aes(x = factor(sample_size), y = avg_sharpe, 
                               color = has_edge, group = has_edge)) +
  geom_line(size = 1) +
  geom_point(size = 3) +
  labs(
    title = "Average Sharpe Ratio by Sample Size",
    x = "Sample Size (Number of Returns)",
    y = "Average Sharpe Ratio",
    color = "Strategy Has\nTrue Edge"
  ) +
  theme_minimal() +
  scale_color_manual(values = c("FALSE" = "blue", "TRUE" = "red"))

# Display plots side by side
gridExtra::grid.arrange(p1, p2, ncol = 2)

The visualisations above provide complementary evidence of how strategy evaluation metrics evolve with increasing sample size:

1. Deflated Sharpe Ratio (DSR) Improves Identification of True Strategies

The left plot clearly demonstrates that the average DSR for strategies with a genuine edge (in red) improves significantly with increased sample size. As we gather more data, genuine strategies become more clearly distinguishable from those without an edge (in blue). Importantly, false strategies do not significantly improve as the sample size increases, instead remaining near a neutral DSR value (~0.5), highlighting DSR’s effectiveness at mitigating false discoveries.

2. Sharpe Ratio Alone Can Be Misleading

Conversely, the right-hand plot shows that the average Sharpe ratios for both true and false strategies converge quickly and closely as the sample size grows. This rapid convergence makes it difficult to differentiate genuine performance from random noise using the Sharpe ratio alone. The substantial drop in Sharpe ratios across both groups as data increases clearly illustrates the Sharpe ratio’s vulnerability to selection bias, particularly in smaller datasets.

Practical Implications

This contrast strongly supports the use of DSR in backtest evaluation, particularly when making high-stakes investment decisions based on strategy performance. The analysis further underscores why industry-standard practice recommends multi-year backtests to reliably distinguish truly profitable strategies from those that merely appear promising due to statistical flukes.

Part 6: Analyzing Performance of Strategies Based on DSR

Now, let’s simulate the performance of strategies selected based on different criteria to see how DSR predicts out-of-sample performance.

# Function to simulate out-of-sample performance
# This provides a realistic test of whether DSR predicts future performance
simulate_oos_performance <- function(n_strategies = 100, 
                                     in_sample_size = 252,
                                     out_sample_size = 252,
                                     edge_pct = 0.05,
                                     edge_size = 0.0005) {
  # Generate in-sample returns (this represents our "backtest" period)
  # We'll use this data to select strategies and calculate DSR
  strategy_data_in <- generate_random_strategies(
    n_strategies = n_strategies,
    n_returns = in_sample_size,
    edge_pct = edge_pct,
    edge_size = edge_size
  )
  
  in_sample <- strategy_data_in$returns
  has_edge <- strategy_data_in$has_edge
  
  # Calculate in-sample Sharpe ratios (what we'd see during strategy development)
  # These are the performance metrics that would guide strategy selection
  in_sample_sharpes <- apply(in_sample, 2, calculate_sharpe)
  
  # Generate out-of-sample returns (this represents future performance)
  # These are completely new random data, simulating what happens post-implementation
  # But strategies with edge still have edge
  out_sample <- matrix(
    rnorm(n_strategies * out_sample_size, mean = 0, sd = 0.01),
    nrow = out_sample_size,
    ncol = n_strategies
  )
  
  # Add edge to the same strategies that had edge in-sample
  for (i in which(has_edge)) {
    out_sample[, i] <- out_sample[, i] + edge_size
  }
  
  # Calculate out-of-sample Sharpe ratios (the "true" performance we care about)
  # This is what would actually be realized when trading the strategy
  out_sample_sharpes <- apply(out_sample, 2, calculate_sharpe)
  
  # Calculate DSR for each strategy based on in-sample data
  # We use this to test whether DSR effectively predicts out-of-sample performance
  dsrs <- numeric(n_strategies)
  for (i in 1:n_strategies) {
    dsr_result <- calculate_dsr(
      in_sample[, i],
      n_trials = n_strategies,  # Number of trials equals number of strategies
      sr_std = sd(in_sample_sharpes)  # Use actual variation in Sharpe ratios
    )
    dsrs[i] <- dsr_result$dsr
  }
  
  # Combine results into a single dataset for analysis
  results <- tibble(
    strategy = 1:n_strategies,
    in_sample_sharpe = in_sample_sharpes,
    out_sample_sharpe = out_sample_sharpes,
    dsr = dsrs,
    has_edge = has_edge
  )
  
  return(results)
}

# Run simulation with 100 strategies
# This gives us enough data to analyze the relationship between DSR and future performance
set.seed(789)  # For reproducibility
performance_results <- simulate_oos_performance(n_strategies = 100)

# Analyze the relationship between in-sample Sharpe, DSR, and out-of-sample Sharpe
# This visualization reveals whether DSR provides useful information about future performance
performance_results %>%
  ggplot(aes(x = in_sample_sharpe, y = out_sample_sharpe, color = dsr, shape = has_edge)) +
  geom_point(size = 3, alpha = 0.7) +
  scale_color_gradient(low = "red", high = "blue") +  # Color by DSR value
  labs(
    title = "Out-of-Sample vs. In-Sample Sharpe Ratio",
    subtitle = "Color indicates Deflated Sharpe Ratio, Shape indicates True Edge",
    x = "In-Sample Sharpe Ratio",
    y = "Out-of-Sample Sharpe Ratio",
    color = "DSR",
    shape = "Has True Edge"
  ) +
  theme_minimal() +
  # Add regression line to show overall relationship
  geom_smooth(method = "lm", se = FALSE, color = "black", linetype = "dashed") +
  # Add reference lines at zero
  geom_hline(yintercept = 0, linetype = "dotted") +
  geom_vline(xintercept = 0, linetype = "dotted") +
  # Add custom shape scale
  scale_shape_manual(values = c("FALSE" = 16, "TRUE" = 17))

# Group strategies by DSR to analyze performance patterns
# This helps us understand whether different DSR ranges predict different outcomes
performance_results <- performance_results %>%
  mutate(dsr_group = cut(dsr, breaks = c(0, 0.2, 0.5, 0.8, 1),
                        labels = c("Very Low (0-0.2)", 
                                   "Low (0.2-0.5)", 
                                   "Medium (0.5-0.8)", 
                                   "High (0.8-1)")))

# Calculate average out-of-sample performance by DSR group
# This quantifies the relationship between DSR category and future returns
dsr_group_performance <- performance_results %>%
  group_by(dsr_group) %>%
  summarise(
    count = n(),  # Number of strategies in each group
    avg_in_sample = mean(in_sample_sharpe),  # Average backtest performance
    avg_out_sample = mean(out_sample_sharpe),  # Average realized performance
    median_out_sample = median(out_sample_sharpe),  # Median (more robust to outliers)
    positive_rate = mean(out_sample_sharpe > 0),  # Proportion with positive returns
    true_edge_rate = mean(has_edge)  # Proportion with true edge
  ) %>%
  arrange(dsr_group)  # Sort by DSR group for readability

# Display the results in a nicely formatted table
kable(dsr_group_performance, 
      caption = "Out-of-Sample Performance by DSR Group",
      digits = 3) %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed"))

Out-of-Sample Performance by DSR Group

dsr_group	count	avg_in_sample	avg_out_sample	median_out_sample	positive_rate	true_edge_rate
Very Low (0-0.2)	53	-0.047	-0.012	-0.018	0.434	0.019
Low (0.2-0.5)	5	0.009	0.015	0.025	0.600	0.000
Medium (0.5-0.8)	4	0.012	-0.019	-0.009	0.500	0.000
High (0.8-1)	38	0.072	0.008	-0.007	0.447	0.105

# Visualize out-of-sample performance by DSR group using boxplots
# This shows the full distribution, not just averages
ggplot(performance_results, aes(x = dsr_group, y = out_sample_sharpe, fill = has_edge)) +
  geom_boxplot(alpha = 0.7) +
  labs(
    title = "Out-of-Sample Sharpe Ratio by DSR Group",
    x = "Deflated Sharpe Ratio Group",
    y = "Out-of-Sample Sharpe Ratio",
    fill = "Has True Edge"
  ) +
  theme_minimal() +
  geom_hline(yintercept = 0, linetype = "dashed", color = "red") +  # Reference line at zero
  scale_fill_manual(values = c("FALSE" = "steelblue", "TRUE" = "darkred"))

Out-of-Sample Validation: Evaluating the Predictive Power of the DSR

The scatterplot above explores a fundamental question in quantitative finance: Does the Deflated Sharpe Ratio (DSR) predict future (out-of-sample) strategy performance?

• Colour indicates DSR: Strategies coloured in blue have higher DSR values (closer to 1), suggesting a high confidence in their genuine predictive ability based on in-sample analysis. Conversely, strategies coloured red (low DSR values) indicate strategies more likely to be statistical artefacts or false discoveries.

• Shape indicates whether a strategy has a true edge: Triangles show strategies deliberately given an actual performance edge. Circles represent purely random strategies without any real edge.

Key Observations:

1. High-DSR Strategies Perform Better Out-of-Sample: The strategies with the highest DSR (blue points) generally cluster in areas with positive out-of-sample performance. This indicates that the DSR effectively identifies strategies that not only performed well historically (in-sample) but also continue to perform well in the future.

2. False Strategies Exhibit Poor Out-of-Sample Performance: Many strategies with low DSR (red points) or those identified as false strategies (circles) show substantial deterioration from in-sample to out-of-sample performance, illustrating a classic regression-to-the-mean effect. These strategies, despite occasionally impressive in-sample results, fail when implemented in live conditions.

3. DSR Provides Incremental Information Beyond Sharpe Ratio Alone: Importantly, the DSR colour gradient reveals that even some strategies with high in-sample Sharpe ratios can have low DSR values, indicating they are less likely to replicate their success out-of-sample. Thus, the DSR serves as an additional layer of quality control beyond the Sharpe ratio.

Practical Implications:

This analysis clearly demonstrates the value of the DSR in strategy evaluation. Practitioners can leverage the DSR to effectively differentiate strategies that have a robust predictive edge from those likely to fail due to selection bias or statistical randomness. Incorporating DSR into strategy selection processes can significantly enhance investment decisions, reducing costly errors from overfitting or selection bias.

Practical Example: Evaluating a Technical Trading Strategy

Let’s evaluate a simple moving average crossover strategy to demonstrate the DSR calculation process in a realistic context.

# Set seed for reproducibility
set.seed(123)

# Generate market data
n_days <- 1000
market_returns <- rnorm(n_days, mean = 0.0004, sd = 0.01)  # ~10% annual return, 16% vol
prices <- 100 * cumprod(1 + market_returns)

# Create strategy: Buy when short MA crosses above long MA, sell when it crosses below
short_window <- 20
long_window <- 50

# Calculate moving averages
short_ma <- zoo::rollapply(prices, short_window, mean, fill = NA)
long_ma <- zoo::rollapply(prices, long_window, mean, fill = NA)

# Generate signals: 1 for long, -1 for short, 0 for no position
signals <- rep(0, n_days)
for (i in (long_window+1):n_days) {
  # Check if we have valid values for comparison
  if (!is.na(short_ma[i]) && !is.na(short_ma[i-1]) && 
      !is.na(long_ma[i]) && !is.na(long_ma[i-1])) {
    
    if (short_ma[i] > long_ma[i] && short_ma[i-1] <= long_ma[i-1]) {
      signals[i] <- 1  # Buy signal
    } else if (short_ma[i] < long_ma[i] && short_ma[i-1] >= long_ma[i-1]) {
      signals[i] <- -1  # Sell signal
    } else {
      signals[i] <- signals[i-1]  # Maintain previous position
    }
  } else {
    signals[i] <- 0  # No signal when we don't have enough data
  }
}

# Convert signals to position vector (1 = long, -1 = short, 0 = neutral)
positions <- rep(0, n_days)
for (i in (long_window+1):n_days) {
  if (signals[i] == 1) positions[i] <- 1
  else if (signals[i] == -1) positions[i] <- -1
  else positions[i] <- positions[i-1]
}

# Calculate strategy returns
strategy_returns <- c(0, positions[-n_days] * market_returns[-1])

# Analyze basic performance
sharpe_standard <- mean(strategy_returns[(long_window+1):n_days]) / 
                    sd(strategy_returns[(long_window+1):n_days]) * sqrt(252)

# Now let's assume this strategy was selected from 50 different parameter combinations
# Calculate DSR
dsr_result <- calculate_dsr(
  strategy_returns[(long_window+1):n_days],
  n_trials = 50,  # 50 different MA combinations were tested
  sr_std = sd(strategy_returns[(long_window+1):n_days]) * sqrt(252)
)

# Display results
cat("Traditional Sharpe Ratio (annualized):", round(sharpe_standard, 2), "\n")

Traditional Sharpe Ratio (annualized): 0.12 

cat("Expected Max Sharpe Ratio from 50 trials:", 
    round(dsr_result$expected_max_sr * sqrt(252), 2), "\n")

Expected Max Sharpe Ratio from 50 trials: 5.69 

cat("Deflated Sharpe Ratio:", round(dsr_result$dsr, 4), "\n")

Deflated Sharpe Ratio: 0 

# Visualize strategy performance
# Create data frame with cumulative returns
performance_data <- data.frame(
  Day = 1:n_days,
  Price = prices,
  ShortMA = short_ma,
  LongMA = long_ma,
  Position = positions,
  CumulativeReturn = cumprod(1 + strategy_returns) - 1,
  MarketCumulativeReturn = cumprod(1 + market_returns) - 1
)

# Plot price and moving averages
p1 <- ggplot(performance_data, aes(x = Day)) +
  geom_line(aes(y = Price), color = "black") +
  geom_line(aes(y = ShortMA), color = "blue", linetype = "solid") +
  geom_line(aes(y = LongMA), color = "red", linetype = "solid") +
  labs(
    title = "Price and Moving Averages",
    x = "Trading Day",
    y = "Price"
  ) +
  theme_minimal()

# Plot strategy cumulative returns
p2 <- ggplot(performance_data, aes(x = Day)) +
  geom_line(aes(y = CumulativeReturn, color = "Strategy")) +
  geom_line(aes(y = MarketCumulativeReturn, color = "Market")) +
  labs(
    title = "Cumulative Returns",
    subtitle = paste("Sharpe:", round(sharpe_standard, 2), 
                    "DSR:", round(dsr_result$dsr, 4)),
    x = "Trading Day",
    y = "Cumulative Return",
    color = "Series"
  ) +
  theme_minimal() +
  scale_color_manual(values = c("Strategy" = "blue", "Market" = "black"))

# Display plots
gridExtra::grid.arrange(p1, p2, ncol = 1)

Practical Illustration of Strategy Evaluation Using DSR

The plots clearly illustrate the pitfalls of relying solely on traditional performance metrics like the Sharpe ratio when evaluating investment strategies:

1. Visual Assessment: The top plot shows typical market price fluctuations and moving averages. Visually, the strategy (buying/selling based on crossovers) seems plausible but generates only random entry and exit signals.

2. Performance Reality Check: In the cumulative returns plot, the strategy’s performance is consistently weaker than the underlying market, despite occasional periods of positive returns. This shows how random strategies can appear superficially promising.

3. Quantitative Confirmation by DSR: While the strategy’s annualised Sharpe ratio (0.12) might appear mildly attractive in isolation, the DSR of 0.00 unequivocally highlights that this performance is illusory. Given that 50 strategies were tested, the observed small positive performance is precisely what we would expect due to selection bias, not genuine predictive power.

Key takeaway:

The Deflated Sharpe Ratio correctly identifies that the modest strategy performance shown here is purely random. This emphasises the critical importance of using robust statistical tools, such as the DSR, in evaluating and implementing trading strategies.

Conclusion

In this tutorial, we have critically explored the issue of selection bias in quantitative investment research, demonstrating how multiple testing inflates performance metrics. Key insights include:

1. False discoveries are common: Even completely random strategies can appear highly profitable when many strategies are tested.
2. Expected maximum Sharpe ratio increases with more trials: Strategies chosen solely because they are the “best performing” from a large set are likely statistical illusions.
3. Deflated Sharpe Ratio (DSR) corrects selection bias: DSR effectively distinguishes genuine strategies from false discoveries by adjusting performance estimates.
4. Sample size critically impacts DSR reliability: Larger samples significantly enhance our ability to correctly identify strategies with true predictive power.
5. DSR predicts out-of-sample performance: High DSR strategies typically demonstrate stronger future performance, validating its practical utility.

This analysis reinforces the need for rigorous statistical approaches, like DSR, to safeguard against backtest overfitting and selection bias, thus improving investment strategy decisions.

Glossary of Key Terms

Backtest Overfitting: Excessive tailoring of a strategy to historical data, capturing noise rather than genuine patterns, resulting in poor future performance.

Deflated Sharpe Ratio (DSR): A performance metric correcting the Sharpe ratio for multiple-testing bias, skewness, kurtosis, and sample size.

False Discovery Rate (FDR): Expected proportion of false-positive findings among strategies identified as profitable.

False Strategy Theorem: Demonstrates the expected inflation in Sharpe ratios purely due to multiple testing.

Multiple Testing Problem: Statistical challenge arising when evaluating numerous strategies simultaneously, increasing the risk of false positives.

Precision: Probability a strategy identified as profitable truly has predictive power.

Recall: Probability a genuinely profitable strategy is identified as profitable in tests.

Selection Bias under Multiple Testing (SBuMT): Performance inflation occurring when only the best-performing strategies are reported among many tested.

Sharpe Ratio: Ratio of excess returns to volatility, commonly used as a measure of risk-adjusted return.

Type I Error (False Positive): Incorrectly concluding a strategy has predictive power.

Type II Error (False Negative): Failing to identify a genuinely profitable strategy.

Exercises for Students

Basic Exercises

1. Sharpe Ratio Distribution: Generate 500 random strategies with 126 days of returns. Plot their Sharpe ratios. Compare these results to 252-day samples. How does reducing sample size affect performance metrics?

2. Expected Maximum Sharpe: Use expected_max_sharpe to plot how expected Sharpe ratios increase with trials: test 10, 50, 100, and 500 trials.

Intermediate Exercises

3. Identifying Strategies with True Edge: Modify generate_random_strategies to introduce genuine edges:

• Evaluate how many true-edge strategies are detected with:
  • Traditional t-tests (5% significance)
  • DSR threshold (> 0.95)

Compare effectiveness of both methods.

4. Monte Carlo Simulation of DSR Effectiveness: Perform Monte Carlo simulations to test:

• Effect of true-edge magnitude
• Impact of varying true-edge proportions
• Role of sample size in strategy detection

Report your findings clearly using visuals.

Advanced Exercises

5. Cross-Validation with DSR: Implement k-fold cross-validation:

• Calculate traditional Sharpe ratio and DSR per fold.
• Compare average in-sample and out-of-sample performance.

6. Portfolio Construction using DSR: Construct portfolios based on:

• Investing only in strategies with DSR > 0.8
• Weighting investments proportionally by DSR
• Equally weighting all strategies

Evaluate out-of-sample performance for each approach.

7. Real-World Strategy Evaluation: Apply DSR to evaluate real strategies:

• Technical indicators (RSI, MACD)
• Factor models (value, momentum)

Assess performance across different market regimes and compare DSR assessment to traditional metrics.

Further Reading

Foundational Papers

• Bailey, D. H., & Lopez de Prado, M. (2014). “The deflated Sharpe ratio.” Journal of Portfolio Management.
• Harvey, C. R., & Liu, Y. (2015). “Backtesting.” Journal of Portfolio Management.
• Bailey, D. H. et al. (2017). “The probability of backtest overfitting.” Journal of Computational Finance.

Practical Implementation

• Lopez de Prado, M. (2018). Advances in Financial Machine Learning. Chapters 10-12.
• Arnott, R. D., Harvey, C. R., & Markowitz, H. (2019). “A backtesting protocol in the era of machine learning.” Journal of Financial Data Science.

Here’s an updated, clear, and structured Online Resources section with verified links for your tutorial:

Online Resources

• Quantitative Finance Stack Exchange (DSR Implementation and Backtesting Discussions) Practical discussions on Deflated Sharpe Ratio implementation and related quantitative finance questions. https://quant.stackexchange.com/questions/tagged/sharpe-ratio
• GitHub Repository: Advances in Financial Machine Learning Exercises and implementations from Marcos Lopez de Prado’s influential book, provided by the community (by BlackArbsCEO). https://github.com/BlackArbsCEO/Adv_Fin_ML_Exercises
• Marcos Lopez de Prado - False Strategy Theorem and Backtest Overfitting (Video) Marcos Lopez de Prado discusses the False Strategy Theorem, backtest overfitting, and strategies to avoid common pitfalls in quantitative research. https://www.youtube.com/watch?v=1ZVyJQ9WqK8

Advanced Topics

• Harvey, C. R., & Liu, Y. (2018). “False (and Missed) Discoveries in Financial Economics.” Journal of Finance.
• Chen, A. Y., & Zimmermann, T. (2022). “Open source cross-sectional asset pricing.” Critical Finance Review.
• Lopez de Prado, M. (2019). “A Data Science Solution to the Multiple-Testing Crisis in Financial Research.” Journal of Financial Data Science.